The Presupposition Problem in Representation Genesis
This paper examines representation genesis—the transition from non-representational physical systems to those with content-manipulable states. It argues that major frameworks in philosophy of mind (Language of Thought, teleosemantics, predictive processing, enactivism, and genetic phenomenology) share a 'Representation Presupposition' structure that prevents them from explaining this first acquisition without circularity. With large language models now achieving high cognitive performance without clear genesis events, the absence of a satisfactory theory becomes urgent.
The paper offers a rigorous conceptual diagnosis establishing that mainstream philosophy of mind faces systematic explanatory deferral when addressing how content-manipulable representation first arises. Its negative argument—that existing frameworks presuppose what they aim to explain—is generally compelling, though the diagnosis depends heavily on accepting the sharp distinction between statistical encoding and genuine representation that the paper stipulates.
The analysis of individual frameworks (especially LOT and teleosemantics) successfully identifies specific 'cognition-side' concepts that presuppose representational organization. The formalization of the Representation Regress in Section 4 provides useful precision, defining CSIDE concepts and showing how explanatory deferral operates across frameworks as diverse as Fodor's nativism and Husserl's genetic phenomenology.
The paper's central distinction between statistical encodings and content-manipulable representations, while philosophically defensible, may strike empirical researchers as under-motivated; the claim that LLMs lack the latter remains contentious and the cited evidence (surface lexical sensitivity, fragility to distribution shifts) is correlational rather than conclusive.
Additionally, the 'regress' identified might be interpreted as ordinary explanatory incompleteness rather than a vicious circularity—a possibility the paper acknowledges but may not fully dispel for skeptics who accept that genesis explanations must stop somewhere.
The evidence consists primarily of conceptual analysis of canonical texts and recent empirical literature on LLM limitations. The paper fairly represents opposing views within enactivism and predictive processing, acknowledging non-representational readings that escape the diagnosis. However, the comparison between frameworks sometimes flattens important differences—teleosemantics' naturalistic grounding differs significantly from LOT's computational architecture—potentially obscuring whether they face the exact same structural problem or merely similar-looking scope limitations.
As a conceptual paper, reproducibility hinges on logical transparency and definitional clarity rather than code or data. The paper succeeds here: it provides explicit three-level distinctions (natural information, statistical encoding, content-manipulable representation), an analytical template for testing frameworks, and formal predicates in Section 4. The minimum adequacy conditions (CC-1 and CC-2) offer falsifiable criteria for future theories. Independent reproduction would require careful tracing of the cited primary texts to verify that the CSIDE concepts are indeed deployed as claimed.
Large language models are the first systems to achieve high cognitive performance without clearly undergoing representation genesis: the transition from a non-representing physical system to one whose states guide behavior in a content-sensitive way. Prior cognitive systems had already made this transition before we could examine it, and philosophy of mind treated genesis as a background condition rather than an explanatory target. LLMs provide a case that does not clearly involve this transition, making the genesis question newly urgent: if genesis did not occur, which cognitive capacities are affected, and why? We currently lack the conceptual resources to answer this. The reason, this paper argues, is structural. Major frameworks in philosophy of mind, including the Language of Thought hypothesis, teleosemantics, predictive processing, enactivism, and genetic phenomenology, share a common feature when applied to the genesis question: at some explanatory step, each deploys concepts whose explanatory purchase depends on the system already being organized as a representer. This pattern, which we call the Representation Presupposition structure, generates systematic explanatory deferral. Attempts to explain the first acquisition of content-manipulable representation within the existing categorical vocabulary import resources from the representational side of the transition itself. We call this the Representation Regress. The paper offers a conceptual diagnosis rather than a new theory, establishing the structure of the problem and deriving two minimum adequacy conditions for any account that avoids this pattern. LLMs make the absence of such a theory consequential rather than merely theoretical.
Pick a starting point or write your own. Challenges run in the background, so you can keep reading while the AI investigates.
No challenges yet. Disagree with the review? Ask the AI to revisit a specific claim.